By Stump Connolly

On the way to school, my son and I sometimes try to imagine the worst job in the world. In a field rife with candidates (algae farmer, manure bagger, spider wrangler), we landed on one that deserves a gold star: fart collector.

There is no equivalent in journalism, but getting assigned to cover the new Pence Advisory Commission on Electoral Integrity that President Trump established to find the 3 million illegal voters he claims participated in the last election has to come close. Because they don’t have much in the way of facts to work with.

A Handful of Examples

Of the 135 million ballots cast last November, The Washington Post found only four instances of voting day fraud in a Nexus-Lexus search of election stories: a woman in Des Moines who voted twice (for Trump); a man in Texas who voted twice because, he claimed, he was a Trump campaign worker testing election security; a Republican election judge in Illinois who filled out an absentee ballot for her recently deceased husband; and a Florida poll worker who filled in absentee ballots for a Miami-Dade mayoral race.

In another survey of local election officials in 42 jurisdictions – where more than 23 million people voted – only 30 cases of suspected non-citizen voting turned up, and secretaries of state across the nation have roundly denounced Trump’s claim.

That has not prevented 15 state legislatures, all under Republican control, from passing tough new laws requiring a photo ID at the polling place. Although Democrats contend these are thinly-veiled attempts to discourage the poor, the elderly, minorities and other left-leaning voters from turning out, the Republican proponents will be flocking to the commission hearings around the country to extol their virtue. But when a final report comes out at the other end, I suspect it won’t pass the smell test.

The Tip of The Iceberg

President Trump has named Mike Pence to be the commission chairman, affixing his name like some “Intel Inside” sticker to authenticate that vote fraud is a real problem, but I suspect he already has his hands full corralling House Republicans, calming allies, excusing Trump’s bad behavior, and, in his spare time, presiding over the Senate.

The real work of the commission will fall to its vice chairman, Kansas Secretary of State Kris Kobach, who has built his political career on making life hard for undocumented immigrants. After winning office in 2011, he championed one of the toughest voter ID laws in the nation, winning unusual powers for a secretary of state to prosecute voter fraud cases. Over the last two years, he has brought nine such cases, but only one involved a non-citizen.

Nevertheless, Kobach has called Kansas “the tip of the iceberg” of fraudulent voting by immigrants. “We’ve had substantial numbers of noncitizens getting on our voter rolls,” he says. “If other states are experiencing the same problem, then I think it would be appropriate for them to consider what Kansas has done.”

Kobach boasted to The New York Times that he has uncovered 125 cases of illegally registered non-citizens in Kansas– out of 1.8 million registered voters – but the Times reporters found only one of the 17 non-citizen voters he discovered in one county had actually cast a ballot; and all nine of his fraud convictions involved people who voted in two states. “Neither a citizenship requirement nor an ID would have prevented those offenses,” The Times wryly concluded.

Meanwhile, a federal court in 2016 struck down the Kansas restrictions on voter registration because they denied more than 18,000 Kansans their constitutional right to vote.

The Evidence

As the presidential campaign came down to the wire, Donald Trump traveled the countryside inveighing against a “rigged” election system. At a rally in Green Bay, Wisconsin, he touted a Washington research paper that proved 14 percent of undocumented immigrants in America were registered to vote.

“You don’t read about this, right? Your politicians don’t tell you about this when they tell you how legitimate all of those elections are,” he said.

The paper by political scientists at Virginia’s Old Dominion University was published in 2014 in Election Studies journal and summarized in a Monkey Cage blog post in The Washington Post titled, “Could Non-citizens Decide the November Election?”

Now that the election is over, The website FiveThirtyEight has published an extensive analysis of the paper, and its role in the election. The 14 percent Trump cited was the high end of a statistical range. There is a 97.5 percent chance the true number is lower, or maybe non-existent. But more importantly, the study is flawed by a common statistical error.

The Methodology

Gulshan Chattha was a Pakistani immigrant doing undergraduate work in political science at Old Dominion when she decided to study the voting habits of naturalized citizens. Since no one asks about naturalization status when you vote, she turned to an often-used political science tool called the Cooperative Congressional Election Study, a national survey administered online to tens of thousands of Americans every year starting in 2006, for her data.

The CCES assembles basic demographic information, including citizenship, political opinions and voting habits. Political scientists trust the CCES because of the size of the sample. But when you break it down into, for instance, the subset of “non-citizens who voted,” the reliability of the numbers breaks down.

In 2008, the CCES surveyed 32,800 people and found 339 identified themselves as non-citizens. In 2010, 55,400 people were surveyed, and 489 said they were non-citizens.

Chattha found that 38 of the 339 non-citizens in 2008 (11.2%) and 13 of the 489 in 2010 (2.7%) also reported that they had voted, a curious fact since non-citizens are not allowed to vote. This is the fact that the Trump camp seized upon to estimate 3 – 5 million immigrants voted against him. But the small size of the sub-sample introduces what social scientists call “a measurement error” that calls into question the validity of the entire paper, Maggie Koerth-Baker contends in her FiveThirtyEight analysis.

Once Trump made her paper the centerpiece of his “rigged election” claim, the author acknowledged the sampling error in her blog. By then, of course, he was president and no one cared.

If, however, the commission wants to plow back into the data, they might also want to look at a new study on the reliability of the CCES data. It found multiple examples of people who claimed to be citizens in 2010 and non-citizens in 2012, or vice versa. These add an additional error rate of 0.1%, which would be enough in 2010, for instance, to say there might not have been any non-citizen voters at all that year.

On Wisconsin

The real life impact of voter ID laws came to prominence last year in Wisconsin, a state that was pivotal to Trump’s Electoral College victory.

The state is hardly a hotbed of voter fraud, but Republican Gov. Scott Walker pushed through a strict voter ID law in 2011 that requires voters to present a driver’s license, state ID, passport, military ID, naturalization papers or tribal ID to cast their ballot. A student ID is acceptable only if it has a signature and two-year expiration date. If you do not have a valid ID, you can cast a provisional ballot that is only counted if you return with proper credentials in three days.

Walker promoted the law as part of a larger voter education effort. “In a society where just about everyone has some form of voter identification, we just need to make sure going forward that we provide it for free,” he said.  But the Brennan Center for Justice says voter IDs for everyone is a pipe dream. In a 2006 survey, the center determined that 21 million voting age citizens in the United States did not have a valid government-issued ID, and another 4.5 million were carrying IDs that, because they had moved or remarried, did not have their current name or address.

Laws Have Consequences

Last week, the Associated Press released an in-depth study of Wisconsin voting in 2016 that indicates the voter ID law probably tipped the state to the Republicans. (Trump beat Clinton by 22,000 votes.)

“By one estimate, 300,000 eligible voters in the state lacked valid photo IDs heading into the election,” the AP’s Christina Cassidy and Ivan Moreno wrote. “it is unknown how many people did not vote because they didn’t have proper identification. But it is not hard to find the Navy veteran whose out-of-state license did not suffice, or the dying woman whose license had expired, or the recent graduate whose student ID was deficient.”

Overall, nearly 3 million people in Wisconsin voted last November. The turnout, however, was 91,000 fewer than in 2012 – and down 41,000 votes in the normally Democratic wards of Milwaukee.

Former Sen. Russ Feingold, who lost his bid in 2016 to regain his seat, calls voter ID laws a scam. “The sole purpose of these laws, including those passed in Wisconsin, is to keep eligible voters from voting, especially minority and low-income voters who tend to vote Democrat.”

But finding new ways to suppress Democratic votes isn’t the only reason we have a Pence Commission on Electoral Integrity. We also need a place for Trump’s bluster to go after all the truth has been squeezed out, and there’s nothing left but the gas that inspired it.

Trackback URL